dollar-cost averaging using the cape ratio: an ... · •did not examine the circumstances of this...

Dollar-Cost Averaging Using the CAPE Ratio: An Identifiable Trend

Influencing Outperformance

By Jon Luskin, MBA CFP®

Define Financial

© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary

Financial Planning Association of MinnesotaOctober 16th, 2017, 2:30 p.m.

Abstract

• Previous research: DCA underperformed LSI most of the time• Did not examine the circumstances of this outperformance

• Concurrent with existing research, found LSI outperformed DCA• 15-year periods• roughly two-thirds of the time • on a nominal return basis, when ignoring taxes and transaction costs

• DCA outperformance a function of CAPE • higher CAPE ratios linked to DCA outperformance

• Time Permitting• What I couldn’t fit into the FPA Journal• Successive Research


Why This Paper


Why DCA & CAPE

• Who uses DCA?

• Over what period?

• When/why?


Why DCA & CAPE

• How to Best Advice Clients with• Business Liquidation

• Real Estate Liquidation

• Pension Lump-Sum Distribution

• Inheritance


Investing in Stocks

• History shows:• Grow money

• In the long term

• Can lose money• In the short term

• Because of volatility

2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016


Dollar Cost Averaging

• The potential for short-term losses may have inspired the idea behind dollar cost averaging

• Don’t invest your money all at once

• Invest small chunks of your money over time

20

00

20

00

20

01

20

01

20

02

20

02

20

03

20

04

20

04

20

05

20

05

20

06

20

07

20

07

20

08

20

08

20

09

20

09

20

10

20

11

20

11

20

12

20

12

20

13

20

14

20

14

20

15

20

15

20

16

20

16

20

00

20

00

20

01

20

01

20

02

20

02

20

03

20

04

20

04

20

05

20

05

20

06

20

07

20

07

20

08

20

08

20

09

20

09

20

10

20

11

20

11

20

12

20

12

20

13

20

14

20

14

20

15

20

15

20

16

20

16

$ $ $$

$$

$ $$ $

$✓© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary


• Proven to be less risky than LSI (as measured by standard deviation)• Williams, R. E., & Bacon, P. W. (1993). Lump Sum Beats Dollar-Cost Averaging.

Journal of Financial Planning, 64-67.

• Dubil, R. (2005). Lifetime Dollar-Cost Averaging: Forget Cost Savings, Think Risk Reduction. The Journal of Financial Planning.



• Proven to be less risky than LSI (as measured by standard deviation)

DCA in progress – less volatile

Stocks

Cash not yet invested

LSI – more volatile



• If less risky, what about a higher investment return?

• Not really

• Less nominal investment return (on average)

20

00

20

00

20

01

20

01

20

02

20

02

20

03

20

04

20

04

20

05

20

05

20

06

20

07

20

07

20

08

20

08

20

09

20

09

20

10

20

11

20

11

20

12

20

12

20

13

20

14

20

14

20

15

20

15

20

16

20

16

20

00

20

00

20

01

20

01

20

02

20

02

20

03

20

04

20

04

20

05

20

05

20

06

20

07

20

07

20

08

20

08

20

09

20

09

20

10

20

11

20

11

20

12

20

12

20

13

20

14

20

14

20

15

20

15

20

16

20

16

$ $ $$

$$

$ $$ $

$✓© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary


• Risk and return are positively correlated


$4,744

$27,120

$21

$0

$1

$10

$100

$1,000

$10,000

19

26

19

27

19

28

19

30

19

31

19

33

19

34

19

35

19

37

19

38

19

40

19

41

19

43

19

44

19

45

19

47

19

48

19

50

19

51

19

52

19

54

19

55

19

57

19

58

19

60

19

61

19

62

19

64

19

65

19

67

19

68

19

69

19

71

19

72

19

74

19

75

19

77

19

78

19

79

19

81

19

82

19

84

19

85

19

86

19

88

19

89

19

91

19

92

19

94

19

95

19

96

19

98

19

99

20

01

20

02

20

03

20

05

20

06

20

08

20

09

20

11

20

12

20

13

20

15

$1 Invested in...

All Stocks Small Stocks 1M T-Bills

Indices data sourced from DFA Web Returns, CRSP 1-10, CRSP 10, One Month T-Bills (Ibbotson & Morningstar)


• DCA doesn’t usually beats LSI• Greenhut, J. G. (2006). Mathematical Illusion: Why Dollar-Cost Averaging

Does Not Work. The Journal of Financial Planning.

• LSI beats 2/3 of the time• Shtekhman, A., Tasopoulos , C., & Wimmer, B. (2012, July). Dollar-cost

averaging just means taking more risk later. Retrieved from Vanguard Group• 1926 to 2011; 6, 12, 18, 24, 30, or 36 months; U.S., U.K. & Australia

• Williams, R. E., & Bacon, P. W. (1993). Lump Sum Beats Dollar-Cost Averaging. Journal of Financial Planning, 64-67• 1926 to 1991; 12-month holding periods; SP & 500



• Because on average, the market move up (upward trending)

• Better off lump-sum investing (LSI)

2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016


1/3 of the time?

• If LSI outperforms, how does DCA outperform during that 1/3 of time?• Flat, downward trending, or volatile markets

• Shtekhman, et al., 2012• Greenhut, 2006

• Previous literature briefly mentioned high volatility made for DCA success• But stopped short of making a full analysis

• Would it be possible to determine what circumstances make for the success of DCA?

• Is it possible to predict when DCA would the better strategy?

• How would we determine this?


CAPE ratio

• Cyclically Adjusted Price to Earnings ratio• AKA Shiller P/E• Invented by economist & Nobel Laurette Robert Shiller

• A valuation metric• Measures if stocks are “cheap” or “expensive”• Looks at company earnings• Over 10 years• Adjusted for price• Relative to stock price• 10 years earnings, adjusted for inflation ÷ stock share price

• Varies between 5+ and 44+• ~16 being average


CAPE ratio

• Has predictive power for investment return

• Significant negative correlation (-0.41), 1871-2015• Shiller, R. (2016, July 28). Online Data - Robert Shiller. Retrieved from Yale Department of Economics: http://www.econ.yale.edu/~shiller/data.htm

-8%

-3%

3%

8%

13%

18%

4 9 14 19 24 29 34 39 44

An

nu

aliz

ed P

rice

A

pp

reci

atio

n

CAPE at begining of 15-Year Period© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary

Using Valuation Metrics as an Investment Strategy• 5-year normalized P/E ratio to tilt equity allocations

• Excess investment returns at highest lowest decile of valuations• Kitces, M., Solow, K., & Locatelli, S. (2011). Improving Risk-Adjusted Returns Using

Market-Valuation-Based Tactical Asset Allocation Strategies. Journal of Financial Planning, 48.

• If a multi-year, inflation adjusted P/E ratio can be used to tilt equity allocation, why not for DCA?• It’s the same thing!


2.50

7.50

12.50

17.50

22.50

27.50

32.50

37.50

42.50

18

81

.01

18

83

.02

18

85

.03

18

87

.04

18

89

.05

18

91

.06

18

93

.07

18

95

.08

18

97

.09

18

99

.1

19

01

.11

19

03

.12

19

06

.01

19

08

.02

19

10

.03

19

12

.04

19

14

.05

19

16

.06

19

18

.07

19

20

.08

19

22

.09

19

24

.1

19

26

.11

19

28

.12

19

31

.01

19

33

.02

19

35

.03

19

37

.04

19

39

.05

19

41

.06

19

43

.07

19

45

.08

19

47

.09

19

49

.1

19

51

.11

19

53

.12

19

56

.01

19

58

.02

19

60

.03

19

62

.04

19

64

.05

19

66

.06

19

68

.07

19

70

.08

19

72

.09

19

74

.1

19

76

.11

19

78

.12

19

81

.01

19

83

.02

19

85

.03

19

87

.04

19

89

.05

19

91

.06

19

93

.07

19

95

.08

19

97

.09

19

99

.1

20

01

.11

20

03

.12

20

06

.01

20

08

.02

20

10

.03

20

12

.04

20

14

.05

Tilting Equity Allocation Relative to CAPE

CAPE

CAPE Average

Underweight equities

Overweight equitiesOverweight equities

Underweight equities


Challenges Using CAPE

• Imperfect valuation metric

• 10 years may be longer than a business cycle

• Technique for measuring inflation has changed with time

• Accounting standards and corporate taxation have changed over time• Wilcox, S. E. (2011, September). A Cautionary Note About Robert Shiller’s

CAPE. Retrieved from AAII: The American Association of Individual Investors


Hypothesis

Can market valuations (CAPE) be used to determine an identifiable trend behind the best opportunities to implement DCA?

5

10

15

20

25

30

35

40

451

88

11

88

31

88

61

88

91

89

11

89

41

89

71

89

91

90

21

90

51

90

71

91

01

91

31

91

51

91

81

92

11

92

31

92

61

92

91

93

11

93

41

93

71

93

91

94

21

94

51

94

71

95

01

95

31

95

51

95

81

96

11

96

31

96

61

96

91

97

11

97

41

97

71

97

91

98

21

98

51

98

71

99

01

99

31

99

51

99

82

00

12

00

32

00

62

00

92

01

12

01

4

CA

PE

Year

1881-2016 1950-2015


Data & Methodology

• S&P 500, Total Return• Yahoo Finance

• 90 Day T-Bills• Board of Governors of the Federal Reserve, 2016

• 1950-2015

• CAPE data from Shiller’s site at Yale

• 15-year rolling time periods• Robust results, vis-à-vis 5 or 10 year• Monthly deposits

• 180 deposits in total• 12 months X 15 years = 180 deposits

• Uninvested cash grew at the risk-free rate© Jon M. Luskin, MBA, CFP® - Confidential & Proprietary

Data & Methodology

• No consideration for taxes or fees• Account fees

• Expense Ratios

• Transaction / Trade Fees / Commissions

• IRA, at Vanguard, using Vanguard funds, etc.


36+ 35.99-31 30.99-26 25.99-21 20.99-16 15.99-11 10.99-6 5.99-0

5%

2%1%

14%

33%

25%

20%

0%

2%1%

4%

14%

29%

31%

18%

2%

Freq

uen

cy o

f O

ccu

rren

ce, %

of

Mo

nth

s w

ith

CA

PE

Per

Tim

e Pe

rio

d

CAPE Ratio

1950-2015 1871-2015


Data & Methodology

• Sample Representativeness

• More frequent higher valuations

• Less lower valuations

Results

• DCA outperformance as CAPE increased• With exceptions

• correlation of 0.43

-6%

-4%

-2%

0%

2%

4%

5 10 15 20 25 30 35 40 45

DC

A O

utp

erfo

rman

ce, %

Starting CAPE for the 15-Year Investment Period


Results

• DCA outperformed 1/3 of the time• As per previous literature

• 1/3 of time, CAPE valuations are above ~18.6

• Using DCA at valuations above ~18.6 averaged an excess return of 0.45 BPS per year• Over 15 years

• Increasing valuations made for a greater degree of DCA outperformance – with the tech bubble exception


Results

• Using DCA at higher valuations averaged higher outperformance• But a non-linear relationship, from the tech bubble

-3.0%

-2.0%

-1.0%

0.0%

1.0%

2.0%

0%

20%

40%

60%

80%

100%Frequency of DCAoutperform

Degree of DCAoutperformance

Linear (Frequency ofDCA outperform)

Linear (Degree of DCAoutperformance)


Results

• Higher CAPE usually meant greater DCA outperformance

• With the tech bubble as an exception

-6%

-4%

-2%

0%

2%

4%

5 10 15 20 25 30 35 40 45

DC

A O

utp

erfo

rman

ce

starting CAPE

DCA outperform

mid 90s


Considerations for Applications

• Perspective #1: You Can Use CAPE to Indicate when to Use DCA• CAPE of 31 – as of Friday, October 13th, 2017

• Will valuations peak at ~44, as per the tech bubble?• Or will valuations surpass ~44?

• Is this the valuation peak?

• Perspective #2: You Cannot Use CAPE to Indicate when to Use DCA• Allocate to a Portfolio that is Always Risk Appropriate for your Client

• “Taking Risk Later”

• What Happens During a Drawdown at Year 16?


Future Research

• Closely Examine Additional Time Periods• 12 months, 24 months, 36 months, 60 months

• Incorporate Taxes & Fees

• Examine Time Periods Back to 1926, 1871• Great Depression

• Examine five- & one-year P/E valuation metrics


What I Couldn’t Squeeze into the FPA Journal

• Best Month for LSI? November • foreign and domestic indices; 1970–1998; 12-month periods

• Atra, R. J., & Mann, T. L. (2001). Dollar-Cost Averaging and Seasonality: Some International Evidence. Journal of Financial Planning, 98–105.

• November has the second highest CAPE ratio, on average, behind December (this study)• Returns on short timelines (one year) determined by momentum

• Return on long timelines (15 years) determined by valuation



Timeline

LongShort

Valuations

High

Low

LSI

LSIDCA

DCA



• Risk-Adjusted Return• LSI ~0.1+ Sharpe ratios, on average

• Shtekhman, et al. (2012)

• 1926 to 2011; 6, 12, 18, 24, 30, or 36 months; U.S., U.K. & Australia

• LSI ~0.07+ Sharpe ratios• Leggio, K. B., & Lien, D. (2003, January). Comparing Alternative Investment Strategies

Using Risk-Adjusted Performance Measures. The Journal of Financial Planning.

• One-year periods; 1926-1999; S&P 500



• LSI outperformed roughly 1/2 of the time

(0.10)

(0.05)

-

0.05

0.10

5 10 15 20 25 30 35 40 45DC

A O

utp

erfo

rman

ce, S

har

pe

Starting CAPE

DCA outperformance

Poly. (DCA outperformance)



• LSI outperformed roughly 1/2 of the time• DCA outperforming when valuations are high and low

• but not median

• DCA strategy outperforms (nominal & risk-adjusted) when valuations are high• Higher Return for Less Risk

• Hence higher Sharpe Ratio



• DCA outperforms on a risk-adjusted basis when valuations are low• Because of consistently high variations in investment return for LSI

• from upward deviations

• Hence higher Sharpe

• But who really cares about upward volatility?

• Sharpe perhaps not an appropriate metric in this circumstance


Postscript: Shorter timelines?

• Shorter timelines of DCA implementation show no consistent result• 6 months, 9 months, 12 months, 18 months


-80%

-60%

-40%

-20%

0%

20%

40%

5 10 15 20 25 30 35 40 45

ou

tper

form

ance

of

DC

A

CAPE ratio at inception

CAPE, 3 mo look back (real)


• Shorter timelines of DCA implementation show no consistent result• 6 months, 9 months, 12 months, 18 months



• CAPE most predictive on longer timelines


(0.5)

(0.6)

(0.7)

(0.6)

(0.5)

(0.4)

(0.3)

(0.2)

(0.1)

-

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

Co

rre

lati

on

to

An

nu

al R

eal

Ret

urn

Length of Investments, Year

Postscript: CAPE is Broken

• Accounting standards changed, making CAPE today different (Seigel, 2016)• National Income and

Product Accounts (NIPA)


0

0.5

1

1.5

2

2.5

3

18

81

18

83

18

84

18

86

18

88

18

90

18

92

18

94

18

95

18

97

18

99

19

01

19

03

19

05

19

06

19

08

19

10

19

12

19

14

19

16

19

17

19

19

19

21

19

23

19

25

19

27

19

28

19

30

19

32

19

34

19

36

19

38

19

39

19

41

19

43

19

45

19

47

19

49

19

50

19

52

19

54

19

56

19

58

19

60

19

61

19

63

19

65

19

67

19

69

19

71

19

72

19

74

19

76

19

78

19

80

19

82

19

83

19

85

19

87

19

89

19

91

19

93

19

94

19

96

19

98

20

00

20

02

20

04

20

05

20

07

20

09

20

11

20

13

20

15

20

16

rolling CAPE using NIPA, deviation from mean, 3 mos lookback

rolling CAPE using S&P, deviation from mean, 3 mos lookback



• w/ NIPA data, a long-enough time period claims below-average valuations



• Incorporating real bond yield can improve forecasting (Vanguard, 2017)

Successive Research


• Testing this same strategy• Using NIPA data

• Vanguard’s algorithm incorporating real bond yields

By Jon Luskin, MBA CFP®

Define Financial

750 B Street, Suite 1750, San Diego, CA 92126

619.577.4002

[email protected]

www.JonLuskin.com/DCA


FPA MN Coordinator: Bonnie Stanley, Financial Planning Association of MN, 3900 Main Street N.E., Columbia Heights, MN 55421, [email protected], 763-781-1212

mailto:[email protected]

http://www.jonluskin.com/

dollar-cost averaging using the cape ratio: an ... · •did not examine the circumstances of this...

Documents